GASTRO library II: Exploring Chemical Bimodalities in Disk Galaxies with GSE-like Mergers and Massive Star-forming Clumps

TL;DR

This study investigates the chemodynamical origin of the Milky Way’s disk bimodality by combining a GASTRO library of SPH+N-body simulations with a Gaia-Sausage-Enceladus–like merger and varying clump formation. The authors show that a brief, significant drop in star formation—driven either by a retrograde GSE-like merger or by an early clumpy SF phase—naturally yields a split in the $[O/Fe]$–$[Fe/H]$ plane, while prograde mergers in non-clumpy disks fail to produce a clear bimodality. A key result is that stars formed in the inner disk can amplify, but do not by themselves create, the outer-disk bimodality, with radial migration further modulating the signature; clumpy disks also generate a non-negligible old $\alpha$-poor population. The findings imply that high-$z$ clumpy galaxies are plausible MW progenitors and that the MW’s bimodality reflects a combination of clump-driven star formation and a single GSE-like event, rather than a simple halt in star formation.

Abstract

We use several smoothed particle hydrodynamics+N-body models as part of the GASTRO library to study the role of high-density star-forming clumpy regions and a single merger on the formation of the $α$-rich and $α$-poor populations in the disk galaxies. These experiments are tailored to mimic what is expected to be the Gaia-Sausage/Enceladus (GSE) accretion event, which occurred circa 10 Gyr ago in the Milky Way (MW). We find that either an early clumpy phase or a retrograde merger significantly reduces the star formation rate (SFR) of the disk, giving rise to a chemical bimodality qualitatively similar to the MW's. The decrease of the SFR as the cause of the chemical bimodality is consistent with previous idealized and cosmological simulations. On the other hand, a prograde radial merger does not significantly modify the SFR of the disk, resulting in no clear chemical bimodality. We further show that stars originating from the inner regions ($R_{form}<4$ kpc) do not create the disk's chemical bimodality, although they can enhance it. Finally, only the models with an early clumpy phase can produce a significant fraction of old, age $>11$ Gyr, $α-$poor stars with disk-like orbits, similar to what has been recently observed in the MW. Our results strengthen the case of clumpy disky galaxies observed at redshift $z\approx 1-2$ as likely progenitors of our Galaxy.

Figures (8)

• Figure 1: Top: The orbital evolution of the satellite galaxy for all the merger models. Low and high supernova feedback models are shown by the green and purple lines, respectively. Prograde and retrograde mergers are represented by dashed and solid lines, respectively. While the subgrid physics and nature of the merger do not influence the orbital evolution, the initial orbital circularity does. More radial orbits (dark lines), $\eta=0.3$, have two apocenters, and the pericenters happen at earlier times compared to the merger on a more circular orbit, $\eta=0.5$. Bottom: The clump mass fraction evolution. The clumps, which only form in the low feedback regime, correspond to 10-20% of the total disk mass in the first 0.5 Gyr; they cease to contribute at around $t=3$ Gyr.
• Figure 2: Face-on view of the stellar density distribution in model c.r.c03. Time increases from left to right and is measured since the start of the simulation at $t=0$ Gyr. The clumpy features are visible early in the evolution of the main galaxy and last up to approximately 3 Gyr. At 1.6 and 3.0 Gyr, the GSE-like dwarf is at its first and nearing its last pericenter, respectively.
• Figure 3: Density in the [$\alpha$/Fe]-[Fe/H] plane for the spatial range $4 < R/{\rm kpc} < 12$, and $|z|<3\, {\rm kpc}$. All models share the same initial gas and dark matter distribution for the MW-like and dwarf galaxies. Clumpy and non-clumpy models are distinguished by titles colored in green and purple, respectively. Isolated, retrograde, and prograde models are shown in the first, second, and third rows, respectively. The satellite's initial orbit circularity is set to 0.3 (second and third columns) or 0.5 (first and fourth columns). All the clumpy models, including the isolated one, develop a chemical bimodality in the disk. However, the prograde mergers do not create the bimodal chemical disk in the non-clumpy model, similar to the non-clumpy isolated model. In non-clumpy models, the bimodality arises primarily from the gas-rich merger, whereas in clumpy models, it is predominantly driven by star formation in the clumps. The black dashed lines show the [Fe/H] interval on which the [O/Fe] probability density function was built (see Figure \ref{['fig:ofe1d']}) to separate both populations. The black solid horizontal line indicates the [O/Fe], which distinguishes the $\alpha-$populations.
• Figure 4: [O/Fe] histogram for the interval $-0.7< {\rm [Fe/H]} < -0.2$, $4 < R/{\rm kpc} < 12$, and $|z|<3\, {\rm kpc}$. All models share the same initial gas and dark matter distribution for the MW-like and dwarf galaxies. Clumpy and non-clumpy models are distinguished by titles colored in green and purple, respectively. Retrograde and prograde models are shown in the second and third rows, respectively. The satellite's initial orbital circularity is set to 0.3 (second and third columns) or 0.5 (first and fourth columns). Dotted lines correspond to the isolated control models shown in the top row. We also include an isolated model with a stronger clumpy phase ("isolated FB10") shown as the dashed line. While all the retrograde merger models can produce a double-peak distribution in [O/Fe], the non-clumpy models with a prograde merger have a similar unimodal distribution compared to their isolated counterpart. Moreover, the clumpy+prograde merger has a smaller fraction of $\alpha-$rich stars, compared to the clumpy isolated model. The initial orbital circularity of the satellite, $\eta$, is indicated in each panel. The dark (light) shaded area corresponds to the contribution of stars born in the inner (outer) parts of the galaxy, i.e. $R < 4$ kpc ($R > 4$ kpc).
• Figure 5: The SFR of the disk, $4 < R/{\rm kpc} < 12$, and $|z|<3\, {\rm kpc}$. The solid line shows the case for the isolated models (top row), and for those with a retrograde (middle row) or prograde merger (bottom row). Purple and green titles correspond to non-clumpy and clumpy models. The contribution to the SFR from the $\alpha-$rich and $\alpha-$poor populations is represented by orange and blue areas, respectively. The models with a clear chemical bimodality have two main features: i) a significant decrease in SFR and ii) the SFR of the $\alpha-$rich stars at earlier times is higher than the $\alpha-$poor population at later times. These occur in the isolated clumpy models and the non-clumpy and clumpy models with a retrograde merger. A prograde merger, however, does not significantly reduce the $\alpha-$poor SFR, and fails to produce the chemical bimodality of the clumpy model. The vertical dashed lines indicate the time of the dwarf's pericenter passages during the merger. The initial orbital circularity of the satellite, $\eta$, is indicated in each panel. The green and purple dashed line shows the SFR for the isolated clumpy and non-clumpy models, respectively.